Bug 1908562 - Pod readiness is not being observed in real world cases
Summary: Pod readiness is not being observed in real world cases
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Node
Version: 4.7
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: 4.7.0
Assignee: Clayton Coleman
QA Contact: Weinan Liu
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2020-12-17 03:02 UTC by Clayton Coleman
Modified: 2021-02-24 15:46 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-02-24 15:45:40 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github openshift cluster-monitoring-operator pull 1020 0 None closed Bug 1908562: Add a recording rule measuring pod readiness 2021-01-19 15:35:26 UTC
Red Hat Product Errata RHSA-2020:5633 0 None None None 2021-02-24 15:46:15 UTC

Description Clayton Coleman 2020-12-17 03:02:34 UTC
The readiness of a pod is a leading indicator of the vast majority of Kube related failures that may impact a workload.  Readiness is a prerequisite for network serving, encodes latency to start, and is impacted by regressions in aggregate behavior.  Since we wish to improve our measurement of the disruption and readiness of both our workload (openshift core) and customer workload during upgrades, the before, during, and after state of readiness can provide a signal for total aggregate health that should in general remain high during an upgrade.

Calculate from a recording rule a metric with one series per pending or running pod which reports 0 if the pod is unready and 1 if the pod is ready. Terminal pods are excluded. Average the pod readiness for openshift-* and !openshift-* workloads and report that to telemetry.

If this metric returns a sufficiently useful indicator for openshift-* workloads, we may decide to investigate clusters that experience a dip immediately after an upgrade.  Large clusters may report a lower average readiness, but would also tend to signal widespread node issues faster.  We may also be able to institute a per component error burn (unready pods is bad as a rule) once we have quantified numbers.

Comment 5 errata-xmlrpc 2021-02-24 15:45:40 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2020:5633


Note You need to log in before you can comment on or make changes to this bug.